Wednesday, March 2, 2011

Formal Complaint Officially Sent To ICGA About Rybka

Something just hit the fan and it's not pretty.

I had been quiety monitoring the Talk Chess forums for sometime now and the it appears that the dam has finally broken.

The creator of Fruit, Fabien Letouzey has filed an official complaint to ICGA complaining that Rybka has taken code from Fruit in direct violation of the GPL. Instead of reproducing the letter, I've provided the link to it. You can see it here from ChessVibes .

I had originally heard about it on the grapevine and was waiting for some official confirmation but it looks like ChessVibes was the first to report it.

What has given enormous weight is the list of whos-who of programmers who put their names on the letter. To those who are unfamiliar with the names, allow me to give a quick blurb for all of them.

Fabien Letouzey - creator of Fruit
Zach Wegner - co-developed Rondo
Mark Uniacke - creator of HIARCS
Stefan Meyer-Kahlen - developer of Shredder
Ed Schröder - creator of Rebel
Don Dailey - a computer chess engine developer since the 90s
Christophe Theron - programmer of Tiger
Richard Pijl - programmer of The Baron
Amir Ban - co-author of Junior
Anthony Cozzie - author of Zappa
Tord Romstad - needs no introduction
Ralf Schäfer - author of Spike
Gerd Isenberg - author of IsiChess and HansDamf
Johannes Zwanzger - author of Jonny

This is a very serious charge. Make no mistake, these are highly respected computer chess programmers with years of experience under their belts and when you get these many people putting their names directly accusing Vasik Rajlich (owner of the hugely successful Rybka) of violating the GPL and using code from Fruit, you can be sure that this will cause a major shakeup.

What now remains to be seen is how David Levy, Jaap van den Herik and the ICGA Board will respond to this serious charge.


  1. I have always preferred older chess programs and distrusted the newer ones. My intuition tells me that some of these newer and popular programs probably copy from those that have been around for years or decades.

    This is why I don't buy the new stuff, I think it's all BS that some genius invents a better algorithm overnight. If anything, in the days of yore, with expensive smaller hardware capacity, the algorithms should have been better.

  2. Hey Linuxguy,

    I think that is the point of the controversy: "No one invents a better algorithm overnight."

    An engine like Shredder or Hiarcs has been around a LONG time and has shown constant improvement.

    I think Shredder , Hiarcs and Junior are the best engines for analysis and sparring partners.

  3. Hi Linuxguy/Tommyg

    Thank u for your comments.

    What was disconcerting was that Rybka in its early days was a very poor chess engine.

    It was only after the release of Fruit 2.1 that some 6++ months later, Rybka jumped right to the front of leading chess engines.

    It is extremely suspicious. Engines don't suddenly experience such large jumps unless the code was rewritten and the thinking and methodolgy changed completely.

    This amounts to a total redesign of the engine and recoding. What the analysis has shown is that Rybka was still using the same bitboard reps as Fruit in their code.

    Vasik has always maintained that Rybka does not copy code from Fruit (if he did, then Rybka would need to release their source as stated in thh GPL) but if this is proven to be false, penalties for Rybka could include being dismissed as a legit chess engine (to join the ranks of Hippolito and Strelka) and possibly barred from all future Computer Chess matches.

    As you stated Tommyg, "No one invents a better algorithm overnight."


  4. I only have a free version

  5. Hi Chess tactics,

    That's ok. I would suggest you get Houdini or at least Stockfish. Even Crafty is extremely strong.

    Fritz is always fantastic because it seems to calculate extremely fast.

    As always, YMMV. :)